The Internet of Things (IoT) is envisioned to become a driving force in the evolution of fifth-generation (5G) mobile networks, autonomous continuous monitoring and control platforms, and low-power consumption devices. Conventionally, these devices are battery-operated. The battery lifespan degrades faster in the case of continuous monitoring devices. Limited battery lifetime motivates us to investigate an eco-friendly solution to solve the issue of limited battery life and frequent replacement. Energy harvesting (EH) is a key-enabling technique that provides a viable solution to the challenge at hand. EH minimizes battery dependence by collecting energy from ambient sources. Although several studies have been conducted on EH-IoT networks, a tutorial on a possible mapping between the use cases/application to the IoT devices and the IoT devices to EH power source still needs to be included. To address this gap, we classify IoT devices based on their applications. Based on the different interfacing sections of layer architecture design of IoT device, this paper tries to fill the gap between IoT applications and the self-sustainable IoT system design. Next, each interfacing section is categorized into layer architecture of IoT device to study layer-specific power requirements (or demand) versus EH sources (or supply). We have considered various applications like agriculture, healthcare, industry, city, security, transportation, and water management. Further, we discuss different energy sources to harness ambient energy and the method to store the harvested energy. Then, the widely used standards and the various steps taken for the interoperability and compatibility of EH-IoT are outlined to understand the elements of the EH interface. Finally, we point out future research issues requiring specific attention for deploying self-sustainable IoT-based delay-limited applications.